This new report, co-authored by Prof. Catherine Banet, Prof. Michael Pollitt, Andrei Covatariu, and Dr. Daniel Duma, looks into the nature and drivers of data centre (DC) energy consumption, Europe’s attractiveness for DC investment, and what net zero energy policies imply for DCs. It also provides regulatory and policy recommendations to ensure adequate grid management and optimisation in the context of growing computing needs.
Understanding Data Centre Energy Consumption and Carbon Footprint Dynamics
A key part of the ICT sector, DCs are a small but growing part of the total EU electricity demand (estimated at 2.7% in 2018). Some projections predict a significant rise by 2030, especially in the context of falling total demand for overall energy but rising electrification. However, the wide range of demand estimates, as identified by the authors, further illustrates the need for more transparency around the underlying data on which current estimates are based.
The report shows how the rise in DC electricity demand has been rather modest against a backdrop of sharply rising demand for data and the growth of large-scale DCs. This can be credited to significant efficiency improvements; in part thanks to more efficient cloud hyperscale DCs. DC owners have also invested considerably in renewable power purchase agreements (PPAs) to cover the equivalent of their electricity demand with low-carbon electricity.
Effectively measuring DCs energy efficiency is a significant challenge, particularly when it comes to defining harmonised criteria that adequately consider all aspects of energy efficiency. The report looks into widely used and relevant metrics. It also considers how DCs’ uninterruptible power supply (UPS) and back-up generation capacity are used.
“Although digital technologies represent an obvious asset in the face of the 21st century environmental challenges, they cannot escape the constraints on energy demand that weigh on every sector of activity on the path to net zero.” Michael Pollitt
Qualification of the Role of Data Centres in the Energy System under EU Law
Like any significant electricity-consuming sector, DCs will be challenged on their consumption and will be expected to interact with the rest of the energy system in a way that facilitates wider decarbonisation. They cannot escape the pricing implications of net zero and its likely requirements for increased differentiation of connection, energy, and ancillary services costs. This presents opportunities for DCs that can offer flexibility to a grid that will be increasingly dominated by variable renewable electricity (VRE).
Europe’s Attractiveness for Data Centre Investments in Light of the Carbon Neutrality Commitment
The authors go on to examine the extent to which Europe attracts DC investment, comparing European cities on criteria such as quality and availability of electrical capacity, temperature, fiscal incentives, and proximity to data demand. This allows a better understanding of why DCs may cluster around particular cities.
They discuss DC markets in Ireland and Denmark to demonstrate the importance of interconnection, renewable electricity shares, and the proximity of heat networks for longer-term prospects in the DC markets that are currently favoured. These case studies illustrate how cluster growth can lead to significant local grid management issues.
Regulatory Proposals Which Could Impact Data Centres
DCs will be increasingly targeted and, in this context, their regulation under EU law must be made consistent across the different legislative acts:
“EU legislative and regulatory initiatives should also aim to foster synergies and mutual benefits from the integration of DCs within the energy system.” Catherine Banet
The report suggests that:
- A ‘dynamic regulatory approach’ should be favoured, based on a mix of legal instruments both legally binding and non-binding nature.
- DCs could serve as role models for the regulation of energy intensive industries, without the need to necessarily be singled out. However, the size of the load may need to be reflected in both connection rules and connection agreements.
- EU regulatory intervention should prioritise the harmonisation of common definitions, principles, obligations, and operating rules.
- One should assess carefully the extended scope of application proposed under the revision of the Network and Information Systems (NIS) Directive (to be replaced by a NIS2 Directive) and the relationship to the European Critical Infrastructure Directive.
- The integration of DCs into high-level energy planning processes should be further pursued.
- The Ecodesign Directive provides a legal basis for the further regulation of the energy consumption of DCs.
- Support for the approach of the new Energy Efficiency Directive (EED), which suggests that all district heating and cooling systems should aim to improve their ability to interact with other parts of the energy system, including DCs.
- Government procurement processes should encourage the use of more environmentally sustainable DCs.
- The revision of the Energy Taxation Directive (ETD) might be an opportunity to harmonise certain practices, at least in terms of minimum harmonisation, to eliminate energy tax breaksthat subsidise higher DC energy consumption.
Conclusions and Recommendations
Over the next ten years, European Internet traffic is expected to grow, but the consequences for DC electricity demand are uncertain. The report highlights emerging issues around some particular cities such as Dublin, with large and growing DC clusters, that may have consequences for the national electricity grid. This has increased grid constraints and could potentially result in increased cost to consumers if there are increased demands on electricity networks and should resource adequacy gaps emerge. It has also resulted in planning permission and siting issues over concerns for the availability of grid connection capacity, as well as political pressure to curtail DC investment in certain locations.
Pressure to better measure energy consumption is already manifesting within existing and proposed EU legislation and can be expected to continue. A wider range of metrics for modelling is needed. Currently information on DC electricity demand, and what drives it, is poor. The industry needs to make more information available so that there can be appropriate independent modelling of the end-to-end energy usage process in ICT. Industry associations and industry standard setting bodies can (and already do) play a key role in spreading best practice on energy consumption and decarbonisation.
There seems little reason to negatively single DCs out in European law, given their need to have UPS and the potential for back-up generation and storage, unlike many other major commercial loads. In places where DC demand is growing sharply, electricity grid issues are related to the addition of any new large load, and hence can be addressed by regulations that target the load characteristics rather than the nature of the load’s economic output. A key test of the efficiency of regulation will be to assess whether it encourages large loads – such as DCs – to facilitate the energy transition and system integration.
Issues encountered locally when introducing DCs in certain locations emphasise the importance of:
- Long term planning and investment into grid infrastructure;
- Visibility of load growth plans from DCs;
- Interconnection and network charges that fully reflect the system costs associated with new, large loads;
- Locational price signals via use of system charges for transmission and distribution; and
- Adequate price signals to encourage flexibility and the co-location of batteries, generation and loads.
The authors conclude that the DC sector has the opportunity to take a lead in corporate citizenship by going beyond the letter of the law and regulation when it comes to innovation in energy reduction and co-operation with others actors. This is already happening.
Finally, the report suggests that DC operators can go further in terms of ensuring that large DCs are capable of contributing to system-wide decarbonisation by appropriate configuration of their UPSs, onsite back up generation, energy storage and energy management to increase grid-level flexibility. This remains an under-researched area for future development.
This report will be presented and debated amongst academics, regulators, industry actors, and policymakers during a CERRE webinar at 14:30 CEST on 13 October 2021.